Vascular leakage or increased vascular permeability in the retina is a common abnormality of diabetic retinopathy and can result in diabetic macular edema (DME), a major cause of vision loss in diabetic patients. Currently, there is no effective drug therapy to treat DME, and DME affects millions of people in the US alone. Over-expression of vascular endothelial growth factor (VEGF) is known to play a pivotal role in the development of vascular leakage in diabetic retinopathy. Angiostatin, a fragment of plasminogen, and kallistatin (or kallikrein-binding protein), a serine proteinase inhibitor, are both potent angiogenic inhibitors. Our previous studies have shown, that angiostatin and kallistatin both significantly reduce vascular leakage in the retina of streptozotocin (STZ)-induced diabetic rat model and in oxygen-induced retinopathy (OIR) rat model. These effects occur at peptide doses much lower than that required for the inhibition of neovascularization, which suggest that these peptides have great therapeutic potential for the treatment of DME. Furthermore, these two angiogenic inhibitors exert their effects on vascular leakage through different mechanisms. Angiostatin is known to interfere with the integrin signaling pathway in endothelial cells, while our preliminary studies showed that the kallistatin effect may be through blocking VEGF binding to VEGF receptors on endothelial cells and thus, blocking VEGF function. As DME is a chronic and multifactoral disorder, we hypothesize that combination of two or more angioeenic inhibitors with different molecular targets may achieve a synergistic effect on vascular leakage and may be more effective for the treatment of diabetic retinopathy. In this propf-of-concept study, we plan to use angiostatin and kallistatin as the first candidates to test this hypothesis and to explore the therapeutic potential of angiogenic inhibitor cocktails. We will determine if angiostatin and kallistatin have synergistic effect on retina vascular leakage in the OIR and STZ-induced diabetes models. This drug cocktail treatment will reduce the dose of each individual drug and thus, minimize possible side effects. This approach could open a new way to treat DME using cocktails of drugs with different molecular mechanisms. [unreadable] [unreadable]